Cancer metastasis is responsible for more than 90% of cancer-related deaths and predicting the location of these secondary tumor sites remains an elusive goal. Studies have demonstrated that more than approximately two-thirds of cancer metastatic sites could be explained by the blood flow pattern between the primary and secondary sites. Development of a precise understanding of cell movement through the vascular system and the likelihood of penetration of the vessel wall is likely critical to achievin the ultimate goal of reliably predicting the vascular regions most likely to incur secondary tumor sites on a per-patient basis. A patient-specific method to predict these patterns will assist in cancer staging, enable identification of unknown primary sites, and inform next-generation treatment therapies that target cancer cells in circulation. We have developed a multiscale computational fluid dynamics model for assessing hemodynamics in image-based arterial geometries, and demonstrated its ability to accurately predict macroscopic quantities related to disease localization and progression. Based on this preliminary data, we hypothesize that (1) cell deformability impacts movement through the vasculature. (2) In vitro measurements can both quantify the range of cell-specific parameters and physiological states that should be used in assessing likely metastatic patterns and validate the computational models. (3) Case-specific simulations can predict likely secondary tumor sites. We propose three specific aims to test these hypotheses: Aim 1. Examine influence of cell deformability on the accurate models of CTC movement, and identify whether the method can be applied at the scale of the full-body. Aim 2. Validate large-scale computational models and predict in vitro measurements of values metastatic sites. Aim 3. Determine the ability of cell-specific computational models of the full-body to predict metastatic patterns observed in vivo. The goal of this application is to develop a method of predicting likely cancer metastasis sites through the use of massively parallel hemodynamic simulations at an unprecedented scale.